ARTIFICIAL INTELLIGENCE TECHNIQUES IN FACILITIES LAYOUT PLANNING: THE DEVELOPMENT OF AN EXPERT SYSTEM (DATABASE)
Abstract
The Facilities layout problem is solved conventionally as a single objective problem. There are not many approaches which treat this problem as a multiobjective problem. The problem is characterized by both quantitative and qualitative information. In this research a multiobjective formulation for the layout problem is given. Expert Systems methodology is suitable when qualitative information becomes necessary to solve a problem. The system developed incorporates rules obtained from the experts. An Expert System is developed to consider the qualitative knowledge. Unlike other Expert Systems this system is characterized by the unique way it generates process knowledge. Procedural rules are used to create the declarative knowledge. The system has modules to generate configurations, based on material flow cost, multi-objective requirements and hard and soft constraints. The system has the capability to learn new process knowledge. The system has the capability to reason and offer explanation. The heuristic methodology is implemented via the Expert System. The system is characterized by the ability to learn new knowledge. The learning is through a dictionary matching. Major segments of the system are implemented in PROLOG. A Facilities layout grammar is proposed for generating alternate configurations. This is based on the concepts of syntactic pattern recognition. This work gives a structure to the domain of Expert System building in manufacturing. An Entity-Relationship decision schema is proposed for knowledge representation in Expert Systems in the manufacturing realm.
Degree
Ph.D.
Subject Area
Industrial engineering|Artificial intelligence
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.